This is intended for use with scripts like addWiki.php to avoid mismatched domain errors.
Bug: T209483
Change-Id: Ie24f83f4e0fad7b01690d2055e1529a4cb0275b5
Non-breaking change. Remaining uses are public interfaces (a constant, two
globals, a config sub-parameter, SQL queries, storage function names), one i18n
message key, and a whole lot of maintenance scripts with calls to the deprecated
function wfWaitForSlaves().
Change-Id: I6ee5ca92ccf6a80c08f53d9efe38ebb4b05064d7
LoadBalancer uses Database::getDomainId() for deciding which keys to use
in the foreign connection handle arrays. This method should reflect any
changes made to the DB selection.
If the query fails, then do not change domain field. This is the sort of
approach that LoadBalancer is expects in openForeignConnection(). Also,
throw an exception when selectDB() fails.
The db/schema/prefix fields of Database no longer exist in favor of just
using the newer currentDomain field.
Also:
* Add IDatabase::selectDomain() method and made selectDB() wrap it.
* Extract the DB name from sqlite files if not explicitly provided.
* Fix inconsistent open() return values from Database subclasses.
* Make a relationSchemaQualifier() method to handle the concern of
omitting schema names in queries. The means that getDomainId() can
still return the right value, rather than confusingly omitt the schema.
* Make RevisionStore::checkDatabaseWikiId() account for the domain schema.
Unlike d2a4d614fc, this does not incorrectly assume the storage is
always for the current wiki domain. Also, LBFactorySingle sets the local
domain so it is defined even in install.php.
* Make RevisionStoreDbTestBase actually set the LoadBalancer local domain.
* Make RevisionTest::testLoadFromTitle() account for the domain schema.
Bug: T193565
Change-Id: I6e51cd54c6da78830b38906b8c46789c79498ab5
So we can make the job voting, preventing other tests from regressing.
These tests can be re-enabled whenever they're made to pass.
Change-Id: I959710138e1e1b37b6ba69561c2920a78208bf12
Previously, if an internal service forwarded the cookies for a
user (e.g. for permissions) but not the User-Agent header or not
the IP address (e.g. XFF), ChronologyProtector could timeout
waiting for a matching writeIndex to appear for the wrong key.
The cookie now tethers the client to the key that holds the
DB positions from their last state-changing request.
Bug: T194403
Bug: T190082
Change-Id: I84f2cbea82532d911cdfed14644008894498813a
Since it takes time for the agent to get the response and set the
cookie and, as well, the time into a request that a LoadBalancer is
initialized varies by many seconds (cookies loaded from the start),
give the cookie a much lower TTL than the DB positions in the stash.
This avoids having to wait for a position with a given cpPosIndex
value, when the position already expired from the store, which is
a waste of time.
Also include the timestamp in "cpPosIndex" cookies to implement
logical expiration in case clients do not expire them correctly.
Bug: T194403
Bug: T190082
Change-Id: I97d8f108dec59c5ccead66432a097cda8ef4a178
Just like 082ed053b6 fixed pre-commit callback errors when new instances
of LoadBalancer are made during that step, do the same for post-commit
callbacks.
Bug: T194308
Change-Id: Ie79e0f22b3aced425cf067d0df6b67e368223e6c
If a pre-commit callback caused a new LoadBalancer object to be created,
that object will be in the "cursory" state rather than the "finalized"
state. If any callbacks run on an LB instance, make LBFactory iterate
over them all again to finalize these new instances.
Make LoadBalancer::finializeMasterChanges allow calls to
already-finalized instances for simplicity.
Bug: T193668
Change-Id: I4493e9571625a350c0a102219081ce090967a4ac
Various revision storage tests assume the sequences are reset
and fail otherwise.
Also disable some failing tests that are not applicable to sqlite
as well as postgres.
Change-Id: Ibdb034121a44e16bb35059a92baafb1867951ea8
This regressed in 14ee3f2, trying to select the server config
"dbname" value as the database on connect. If that config is
bogus, then the connection attempt would fail. Instead, always
pass null to the driver's connection function if the DB domain
doesn't matter.
This affected WMF sites during lag checks, on account of the
"serverTemplate" value in $wgLBFactoryConf always using the local
wiki domain (whether the cluster had such a database or not).
Use strlen() as mysqli sees null and "" as the same for $dbname.
Bug: T186764
Change-Id: I6699d17c0125a08415046211fee7906bbaf2c366
The binlog file/pos where only being used in __toString() for the GTID
case. Make that method use the GTID set instead and avoid querying the
old-fashioned binlog fields all together in that case. The STATUS
queries involve some global lock contention.
Bug: T180918
Change-Id: I18123a702e4f554b87bf5f90017b248062e73049
When LoadBalancer opens new local domain connections, it currently
assumes that the domain specified by the server info array is the
same. For sanity, make sure that the handle is set to the local
domain.
The main LBFactory/LoadBalancer use $wgDBname/$wgDBprefix as the
local domain, corresponding with wfWikiId(). This relation is set
automatically in MWLBFactory. If $wgLBFactoryConf/$wgDBservers is
manually configured in a way breaking this correspondance, then it
is misconfigured.
Fixes made to avoid test failure:
* Make sure LoadBalancer::setDomainPrefix() updates the local
domain alias member. Also do not bother changing the domain of
foreign connections.
* Use the right domain ID for the connection array key names in
LoadBalancer::openForeignConnection().
* Now that JobQueueTest no longer mistakenly uses the non-test
tables, force it to use the main DB_MASTER handle so that it can
see the unit test tables even if they are TEMPORARY; such tables
are tied to the TCP connection, so separate handles see different
temporary tables.
Change-Id: I56f8b32fe957f984b8c9753e6db3b20abe96b038
This handles multi-DB transactions properly, instead of causing wait
timeouts in the WaitConditionLoop. It also is more correct in using
a counter instead of relying on wall clocks.
In addition:
* Refactor related code in MediaWiki.php to be comprehensible.
* Always send the cookie even the "remote wiki redirect" case.
* Renamed ChronologyProtector field and constant to avoid any
confusion of "wait for server X to reach Y" with "wait for Y
to show up in position store".
* Add an "asOfTime" field to the position keys for debugging.
Bug: T182322
Change-Id: I5c73cd07eaf664f02ba00c38fab9f49b609f4284
Change 950cf6016c took care of the most,
but a few remain, either outside of includes/ and maintenance/
directories (which that change was limited to), or in code introduced
afterwards.
Change-Id: I9c363d0219ea7e71cde520faba39406949a36d27
I was bored. What? Don't look at me that way.
I mostly targetted mixed tabs and spaces, but others were not spared.
Note that some of the whitespace changes are inside HTML output,
extended regexps or SQL snippets.
Change-Id: Ie206cc946459f6befcfc2d520e35ad3ea3c0f1e0
The class DatabaseMysql corresponds to the PHP extension 'mysql' which is
deprecated since PHP 5.5, and the class DatabaseMysqli corresponds to the
PHP extension 'mysqli' which is the current MySQL driver. Given the class
DatabaseMysql is likely to be removed in the future, this change is a
first step towards this.
Bug: T120333
Change-Id: I7b8363838449cae09b360221d8f50ace3a008f11
Causes HHVM 3.12.7 to segfault and or:
Fatal error: Stack overflow in includes/libs/objectcache/BagOStuff.php on line 754
This reverts commit b47ce21cec.
Bug: T150833
Change-Id: Ia52fdf65f5e386c0eaa1046328680fb5c8a9081a
This works by setting a callback to return the cache set
options. The callback will watch DB reads and create a
merged result from said usage.
This handles callers that are missing getCacheSetOptions().
Change-Id: Ia264f011e45e8cf105480955dad7e2c4c2357b73
It's no longer an LBFactory subclass (since 5d4b009cf), and the MW-prefix
seems more natural for this class.
The class has only existed since for about 2 weeks (since 0e5cd18b7) and
is not used outside MediaWiki core across Wikimedia Git.
Change-Id: I34be982b5d10ad03e062033da9c40b4a01665289
Fixed some errors that popped up in CI:
* Also cleanup $domain handling in reuseConnection().
* Fix empty string handling in openForeignConnection() where
the empty string check against $dbName failed since an empty
string $domain results in $dbName being null.
Change-Id: Ie78fefa1acb401fe4e8bdc96b75053692aa0a925
* Refactored LBFactory a bit to make this possible.
* Move newChronologyProtector() up to LBFactory and
make a lazy-loading method instead.
* Move appendPreShutdownTimeAsQuery() up to LBFactory.
* Inject the web request values for LBFactory from Setup.php.
* Remove unused laggedSlaveUsed() method.
Change-Id: Ie8a38a6f4d6359680eb6a5be24a34e30b9816479
* Add DatabaseDomain class to handle passing domains around.
It also can be cast to and from strings, which are of the same
format as wfWikiId() except with hyphens escaped.
* Make IDatabase::getDomainID() use these IDs so they can be
passed into LoadBalancer::getConnection() and friends without
breaking on sites with a hyphen in the DB name.
* Add more LBFactory unit tests for domains.
Bug: T145840
Change-Id: Icfed62b251af8cef706a899197c3ccdb730ef4d1
The former extends the later with MW-specific logic.
Also removed a wf* method call from ChronologyProtector.
Change-Id: I325f59b7467ab9c2137731d1ce69816f5a020f03
This lets callers check whether a user recently change a DB and
possibly try harder to reflect certain changes or refresh caches.
Also simplified some logging code a bit.
Change-Id: Ia1168cf0d46cfdee046838ce4c5a6294e4d81760
Before, it just used the immediate master entry, which could be another
slave. In that case, it may not even exists at all.
Bug: T119648
Change-Id: Iea970b81ad2c9855aafcccf0bb0662fc0b3a8d4d
* It now works for users without using sessions.
Sessions should not be cluttered with things
unrelated to authentication and tokens.
* Public services doing api.php requests on behalf
of a users only need to set XFF headers (as normal)
for position wait logic to trigger. They can opt out
of ChronologyProtector via a new HTTP header
"ChronologyProtection: false".
* Requests across subdomains, such as the SUL2 handshake
for CentralAuth on account creation, now have position
wait logic applied. This helps avoid anomolies were a
row just written in the last request may not be seen.
* Use merge() to avoid rolling back master positions if
the user has multiple tabs open and doing writes at once.
* $_SESSION global state is gone from ChronologyProtector.
* Cleaned up post-send LBFactory::shutdown() logic for
avoiding master position writes with an explicit flag.
* Use 'replication' debug log group in more places.
Bug: T111264
Change-Id: Ib25d05994d62b25c2f89e67b7f51009c54f4bca8
* wfGetDB( DB_MASTER )->getLBInfo() now shows the master flag
* A corresponding slave flag was also added
* Added a few badly needed LBFactory/LoadBalancer tests
Change-Id: I9254b12cff63af7d754a3a14c5db44276f58d280
This was used in 2 special classes, the logger classes and spread
to a few other random classes.
Afaik this has no meaning. Is for something we don't use, and
goes against the meaning of '@section' in Doxygen, which we do
use.
In Doxygen output, all LICENSE references became links to
ProfilerXhprof (the one Doxygen encoutered first).
Bug: T72328
Change-Id: Icc7c443245c70bc0f549bee7d105eef5691c864d
Instead test that LBFactory::getLBFactoryClass() does the right thing.
In LBFactory::getLBFactoryClass(), the method now requires a configuration
array as a parameter, and returns the class name as a string.
LBFactory::singleton() now passes in $wgLBFactoryConf into getLBFactoryClass(),
while tests can use test configuration.
Only use of LBFactory::getLBFactoryClass() in core is in LBFactory::singleton().
No extension in gerrit uses getLBFactoryClass() and I can find no use of it
when searching places like Github and Google, so it should be safe to make
such change in LBFactory.
Bug: 59105
Change-Id: I71ae7df16bc5c214b9389125140bca5ce68d274c
